Formation of Nano-Crystalline Todorokite from Biogenic Mn Oxides

Home , Birnessite, Cryptomelane, Manganese oxide, Manganite, Oxide mineral, Psilomelane

Available online at www.sciencedirect.com

Geochimica et Cosmochimica Acta 74 (2010) 3232–3245 www.elsevier.com/locate/gca

Formation of nano-crystalline todorokite from biogenic Mn oxides

Xiong Han Feng a,1, Mengqiang Zhu a, Matthew Ginder-Vogel a,b, Chaoying Ni c, Sanjai J. Parikh a,2, Donald L. Sparks a,b,*

a Environmental Soil Chemistry Research Group, Department of Plant and Soil Sciences and Center for Critical Zone Research, 152 Townsend Hall, University of Delaware, Newark, DE 19716, USA b Delaware Environmental Institute, University of Delaware, Newark, DE 19716, USA c Department of Materials Science and Engineering, 201 Dupont Hall, University of Delaware, Newark, DE 19716, USA

Received 15 May 2009; accepted in revised form 3 March 2010; available online 12 March 2010

Abstract

Todorokite, as one of three main Mn oxide phases present in oceanic Mn nodules and an active MnO6 octahedral molecular sieve (OMS), has garnered much interest; however, its formation pathway in natural systems is not fully understood. Todorokite is widely considered to form from layer structured Mn oxides with hexagonal symmetry, such as vernadite (d-MnO2), which are generally of biogenic origin. However, this geochemical process has not been documented in the environment or demonstrated in the laboratory, except for precursor phases with triclinic symmetry. Here we report on the formation of a nanoscale, todorokite-like phase from biogenic Mn oxides produced by the freshwater bacterium Pseudomonas putida strain GB-1. At long- and short-range structural scales biogenic Mn oxides were transformed to a todorokite-like phase at atmospheric pressure through refluxing. Topotactic transformation was observed during the transformation. Furthermore, the todorokite-like phases formed via refluxing had thin layers along the c* axis and a lack of c* periodicity, making the basal plane undetectable with X-ray diffraction reflection. The proposed pathway of the todorokite-like phase formation is proposed as: hexagonal biogenic Mn oxide ? 10-A˚ triclinic phyllomanganate ? todorokite. These observations provide evidence supporting the possible bio-related origin of natural todorokites and provide important clues for understanding the transformation of biogenic Mn oxides to other Mn oxides in the environment. Additionally this method may be a viable biosynthesis route for porous, nano-crystalline OMS materials for use in practical applications. Ó 2010 Elsevier Ltd. All rights reserved.

1. INTRODUCTION

Mn oxides are environmentally ubiquitous and an important source of reactive mineral surfaces in the envi- * Corresponding author at: Environmental Soil Chemistry ronments. There are over 30 known Mn oxide/hydroxide Research Group, Department of Plant and Soil Sciences and minerals resulting from the numerous environmental Mn Center for Critical Zone Research, 152 Townsend Hall, University oxidation states [Mn(II), Mn(III) and Mn(IV)] and an array of Delaware, Newark, DE 19716, USA. Tel.: +1 302 831 6378; fax: of atomic arrangements (McKenzie, 1989; Dixon and Skin- +1 302 831 0605. ner, 1992; Post, 1999). These minerals participate in a vari- E-mail address: [email protected] (D.L. Sparks). 1 ety of chemical and biological reactions that aﬀect the water Present address: College of Resources and Environment, quality of marine and soil systems (Villalobos et al., 2003; Huazhong Agricultural University, Wuhan 430070, PR China. Tebo et al., 2004; Webb et al., 2005a), and due to their reac- 2 Present address: Department of Land, Air and Water Resources, One Shields Avenue, The University of California, tivity have been called “scavengers of the sea” (Goldberg, Davis, CA 95616, USA. 1954). The basic building block of Mn oxides is the

MnO6 octahedron. These octahedra can be assembled ally possessed a layered topology (Saratovsky et al., through corner and/or edge sharing into a variety of struc- 2006). In the presence of U(VI) the marine bacterium Bacil- tures that fall into two basic categories: (1) layer structures lus sp. strain SG-1 forms poorly ordered Mn oxide tunnel (phyllomanganates) and (2) chain, or tunnel structures (tec- structures, similar to todorokite (Webb et al., 2006); how- tomanganate) (McKenzie, 1989; Dixon and Skinner, 1992; ever, this phase has not been identified in environmental Post, 1999). According to the tunnel size, tectomanganates systems. Synthetic todorokites are generally obtained from can be denoted as T(m n). Todorokite, a family of tunnel modifying a layer structured Mn oxide with triclinic sym- structure Mn oxides with a T(3 3) array of edge-shared metry via a hydrothermal chemical route at relatively high MnO6 octahedra, is commonly associated with ferromanga- temperature and pressure (Golden et al., 1986; Shen et al., nese oxides from marine (Burns and Burns, 1978a; Chukh- 1993; Feng et al., 1995, 1998; Vileno et al., 1998; Ching rov et al., 1979; Mellin and Lei, 1993; Takahashi et al., et al., 1999; Luo et al., 1999; Liu et al., 2005). The forma- 2007) and terrestrial (Turner and Buseck, 1981; McKeown tion of todorokite is greatly accelerated under mild reflux and Post, 2001; Manceau et al., 2007) settings. Mn oxide conditions at atmospheric pressure, enabling the simulation minerals are of potential economic interest because they of formation processes for naturally occurring todorokite are often enriched in Co, Ni, Cu and other strategic metals, (Feng et al., 2004; Cui et al., 2006, 2008, 2009a,b). Here including platinum group and rare earth elements (Post, we describe the transformation of biogenic Mn oxide into 1999; Glasby, 2006). In addition, todorokite has many po- a todorokite-like phase. The transformation products were tential industrial applications, including use as sorbents, characterized using X-ray absorption near edge structure heterogeneous catalysts, sensors, and rechargeable battery (XANES) and extended X-ray absorption fine structure cathodes (Shen et al., 1993; Vileno et al., 1998; Ching (EXAFS) spectroscopies, synchrotron-based X-ray diffrac- et al., 1999; Feng et al., 1999; Suib, 2008; Cui et al., 2009a). tion (SR-XRD), transmission electron microscopy (TEM), Many of these Mn oxides are formed by microbial oxi- field emission gun scanning electron microscopy (FEG-SEM) dation of soluble Mn(II). In fact Mn-oxidizing biota (i.e., and high-resolution transmission electron microscopy bacteria and fungi) are commonly distributed throughout (HR-TEM). We also propose a potential transformation freshwater, ocean, and soil environments and catalyze the pathway and mechanism for biogenic Mn oxides transfor- oxidation of Mn(II) at faster rates than abiotic processes mation into todorokite-like minerals. (Nealson et al., 1988; Takematsu et al., 1988; Tebo et al., 2004). Recent studies characterizing microbial Mn(II) oxi- 2. EXPERIMENTAL METHODS dation products reveal that they are exclusively X-ray amorphous, hexagonal, layer type Mn oxides with nano- 2.1. Biogenic Mn oxide production particle size similar to d-MnO2 (Bargar et al., 2005, 2009; Webb et al., 2005a,b; Miyata et al., 2006; Saratovsky Biogenic Mn oxides were produced by cultures of Pseudo- et al., 2006; Villalobos et al., 2006). Reaction of Mn(II) monas putida strain GB-1, provided by B.M. Tebo (Oregon and/or coexisting ions with the primary biogenic Mn oxide Health and Science University). Bacteria were grown in mineral yields abiotic secondary products, including 10-A˚ 500 mL L. discophora media in 1800 mL Erlenmeyer flasks Na phyllomanganate, feitknechtite, hausmannite and man- at 30 °C and 200 rpm in a temperature-controlled incubator ganite (Mandernack et al., 1995; Bargar et al., 2005). The with an orbital shaker. The Leptothrix media contained occurrence of diverse Mn oxides in surface environments 0.5 g L1 yeast extract and casamino acids, 1 g L1 glucose, may result from secondary products of biogenic Mn oxida- 10 mM HEPES buffer (pH 7.5), 2 mM CaCl2, 3.3 mM tion (Tebo et al., 2004; Villalobos et al., 2003, 2006; Bargar MgSO4, 3.7 lM FeCl3 and 1 mL trace element solution et al., 2005, 2009). (10 mg/L CuSO45H2O, 44 mg/L ZnSO47H2O, 20 mg/L The conversion pathways of biogenic Mn oxides into CoCl26H2O, and 13 mg/L Na2MoO42H2O) (Boogerd and other Mn oxides, especially tunnel structure Mn oxides de Vrind, 1987). Inoculum cultures were prepared by growing (e.g., todorokite), remains poorly understood. Although bacteria from a L. discophora agar plate in MSTG media todorokite is often found associated with Mn oxides of (2 mM (NH4)2SO4, 0.25 mM MgSO4, 0.4 mM CaCl2, microbial origin in ocean nodules, the pathways and mech- 0.15 mM KH2PO4, 0.25 mM Na2HPO4, 10 mM HEPES, anisms of todorokite formation from biogenic Mn oxides in 0.01 mM FeCl3, 0.01 mM EDTA, 1 mM glucose, and 1 mL nature are currently unknown (Burns and Burns, 1978b; trace metal solution) for 12 h at 30 °C(Parikh and Chorover, Siegel and Turner, 1983; Mandernack et al., 1995; Post, 2005). Cells were harvested after 19 h, via centrifugation at 1999; Buatier et al., 2004; Bodei et al., 2007). This is largely 10,000 RCF, at which time the cells have a maximum oxidiz- due to difficulty simulating the geochemical processes in- ing capacity. The harvested cells were rinsed with a solution volved in the mineralogical transformation from layered, of 10 mM HEPES at pH 7 to remove metabolites from the biogenic Mn oxides into tunnel structure Mn oxides. These spent media. The cells harvested from each 500-mL culture difficulties stem from the length of time these processes take were re-suspended in 1 L of autoclaved 50 mM NaCl and at room temperature (Cui et al., 2006); additionally identi- 10 mM HEPES at pH 7. Filter-sterilized MnSO4 solution fication of poorly crystalline phases in a mixed system is was added to the above solution after autoclaving to a final problematic. Mn oxide minerals formed by the freshwater concentration of 100 lM. The suspensions were shaken at Leptothrix discophora SP6 were initially thought to have a 200 rpm at 30 °C for 48 h, at which time the Mn(II) in the todorokite-like tunnel structure (Kim et al., 2003); however, solution, measured by the formaldoxime colorimetric further investigation indicated the biogenic product actu- method (Burle and Kirby-Smith, 1979), was exhausted. The 3234 X.H. Feng et al. / Geochimica et Cosmochimica Acta 74 (2010) 3232–3245 biogenic Mn oxides were then collected by centrifugation at were washed, centrifuged, and re-suspended several times to 3000 RCF. At this centrifugal speed, non-membrane bound remove the loosely bound Mn2+ and Mg2+ remaining from EPS (exopolymer substances) remained in the supernatant the corresponding treatments. After vacuum filtration, the and was then discarded. The biogenic Mn oxides were then XAFS samples were prepared by mounting the wet paste re-suspended in 50 mM NaCl and 10 mM HEPES (pH 7) in thin (1 mm) plastic holders in a 20 5 mm slot which and allowed to settle overnight. The EPS in this supernatant was sealed with Mylar film. Mn K-edge XAFS data were was then removed to further purify the biogenic Mn oxides. collected using beamline X-11A at the National Synchro- This procedure was repeated several times to ensure that only tron Light Source (NSLS), Brookhaven National Labora- trace amounts of EPS remained associated with the biogenic tory (Upton, NY). The electron beam energy was 2.5– Mn oxides. 2.8 GeV, with a maximum beam current of 300 mA. The monochromator consisted of two parallel, channel-cut 2.2. Transformation of the biogenic Mn oxide Si(1 1 1) crystals with a vertical entrance slit opening of 0.5 mm. The beam size on the sample was maintained at After purification as described above, the biogenic Mn 2 10 mm. The samples were mounted 45° to the incident oxide, collected from three 1-L 100 lM Mn(II), 50 mM beam and Mn K-edge EXAFS data were collected over the NaCl and 10 mM HEPES solutions, was mixed before re- energy range 6339–7286 eV in fluorescence mode using an 0 suspension in 250 mL 1 M MgCl2 solution and exchanged Ar filled Lytle detector. An internal reference (Mn ) was 2+ for 12 h. After centrifugation at 10,000 RCF, the Mg - collected concurrently (E0 = 6539 eV) for energy calibra- exchanged biogenic Mn oxide (hereafter BMO-Mg) was tion. The first ionization chamber was filled with 50% N2 re-suspended in a 250 mL 1 M MgCl2 solution at pH 5.1 and 50% He, while the second and third ionization cham- in a 500 mL Erlenmyer flask connected with a glass con- bers were filled with 100% N2.A3lx Cr filter, one to denser cooled by using tap water in the outer jacket. Then two sheets of Al foil, and Soller slits were used to limit the suspension was heated to and kept at reflux under stir- the impact of elastic and Compton radiation as well as filter ring on a combined hot-plate and magnetic-stirrer. Aliquot the fluorescence signal. The fluorescence and transmission suspensions (50-mL) were taken and cooled down to room data of each sample were compared to check for self- temperature at 8 h, 24 h time intervals. After 48 h of reflux, absorption effects, which were not observed. Harmonics the heat was stopped and the residual suspension was were eliminated from the incident beam by detuning the cooled to room temperature. The refluxed solid products, monochromator by 30% of I0. Multiple scans (P3) were BMO-8 h, BMO-24 h and BMO-48 h designated for prod- collected for each sample to improve statistics. Reference ucts refluxed for 8, 24 and 48 h, respectively, were obtained Mn oxide mineral samples were prepared for analysis by via filtering with 0.22 lm filters and then washed with mixing finely ground powder of each mineral with boron ni- 25 mL of distilled deionized water for two times. The pH tride (BN) to 10% Mn by weight. Each sample was then and Mn(II) in the supernatants were determined using a loaded into an individual acrylic sample holder and sealed pH meter and Inductively Coupled Plasma Atomic Emis- with Kapton tape. Non-adhesive Kapton film was used to sion Spectrometry (ICP-AES), respectively. seal the sample cell to avoid any interaction of the sample with the tape adhesive. The Mn K-edge XAFS data were 2.3. Preparation of the reference Mn oxide minerals collected from reference Mn oxide minerals in transmission mode. Reproducibility of the transmission spectra collected Todorokite (hereafter Todorokite-STD) was prepared at several different locations and at a long duration on se- using a previously described reflux method (Feng et al., lected samples confirmed sample homogeneity and no sam- 2004). Cryptomelane was synthesized by modifying the pro- ple damage by the X-ray radiation. All XAFS data cedure of McKenzie (McKenzie, 1971; Feng et al., 2007). reduction and analysis was performed using SIXPack d-MnO2, a disordered hexagonal layer manganate (Villalo- (Webb, 2005). Mn K-edge EXAFS data were fit in R space bos et al., 2003, 2006), was prepared using a “redox” meth- using a full multiple scattering model based on a phyllo- od with stoichiometric amounts of KMnO4 and MnCl2 manganate structure (Webb et al., 2005a). (Gadde and Laitinen, 1974). Acid birnessite (hexagonal birnessite), was prepared by reducing KMnO4 with concen- 2.5. Synchrotron-based X-ray diffraction (SR-XRD) trated HCl at boiling temperature (McKenzie, 1971). Random stacked birnessite (RSB), a disordered triclinic bir- SR-XRD patterns were recorded from wet paste sam- nessite, was synthesized through oxidation of Mn(OH)2 by ples in transmission geometry with a MAR345 image O2 in alkali medium (Yang and Wang, 2002). Triclinic bir- plate at an incident X-ray energy of 12,732 eV nessite was prepared by aging the RSB suspension at 313– (0.9742 A˚ ) at SSRL beamline 11-3. Wet sample slurries 373 K (Yang and Wang, 2002). MnO, Mn2O3 (bixbyite) were placed in an aluminium sample cell between Lexan and MnO2 (pyrolusite) were purchased from Aldrich. The (polycarbonate) windows. Two-dimensional XRD pat- purity of these phases was confirmed by X-ray diffraction. terns were calibrated with lanthanum hexaboride (LaB6) and integrated to one-dimensional patterns with Fit2d 2.4. X-ray absorption fine structure (XAFS) spectroscopy (Hammersley et al., 1996). The background contributions due to the Lexan windows and water in the sample were Prior to XAFS spectroscopic analyses, biogenic Mn removed using XRDbs (http://ssrl.slac.stanford.edu/ oxide, Mg2+ exchanged minerals and the refluxing products ~swebb/xrdbs.htm). Formation of todorokite from biogenic Mn oxides 3235

2.6. FEG-SEM and TEM analysis

FEG-SEM and TEM micrographs were collected at the Bio-imaging Center at Delaware Biotechnology Institute at the University of Delaware. Morphology of the biogenic Mn oxide and the refluxed products were imaged with a Hitachi S-4700 field emission gun scanning electron micro- scope and a Zeiss CEM 902 TEM. Prior to analysis the samples were fixed using glutaraldehyde, postfixed with 1% osmium tetraoxide, and then dehydrated with a series of ethanol–water solutions. After critical point drying, the samples were mounted on double-sided carbon tape and carbon coated for FEG-SEM observation. For TEM analysis the samples were fixed with glutaraldehyde, postfixed with 1% osmium tetraoxide, stained with 0.5% uranyl ace- tate, and then dehydrated with acetone. Then they were Fig. 1. XANES spectra of BMO, BMO-Mg, BMO-8 h, BMO-24 h, embedded in resin and thin sectioned with a glass knife BMO-48 h and Todorokite-STD. The white lines for BMO and on a microtome for TEM observation. BMO-Mg are at around 6562.4 eV, and those for BMO-8 h, BMO- 24 h and BMO-48 h are 6560.9 eV. The white line for the todorokite standard is 6561.5 eV. 2.7. HR-TEM analysis

The HR-TEM analyses were performed on the above TEM samples and sample suspensions air dried on a holey carbon grid with a JEOL JEM 2010 FEF electron micro- scope operated at 200 kV.

3. RESULTS

3.1. XANES and EXAFS spectroscopy

3.1.1. XANES The lineshapes of Mn K-edge XANES spectra are sensitive to changes in oxidation state (peak position) and local coordination environment (peak shape) (Villalobos et al., 2003). XANES was used as the principal method to deter- mine the average oxidation state (AOS) and coordination geometry of Mn in samples maintained under natural conditions (Villalobos et al., 2003, 2006; Webb et al., 2005a). A Fig. 2. Quantification of AOS in BMO, BMO-Mg, BMO-8 h, standard curve was generated using the edge positions of BMO-24 h, BMO-48 h and Todorokite-STD. AOS was determined by location of absorption edge energies (inflection point) in MnOx standards with known oxidation states (Fig. 1). The AOS of Mn in BMO and BMO-Mg was found to be reference compounds of known oxidation states. The edge energies observed in the XANES of MnO, Mn O (bixbyite) and MnO 3.8 ± 0.3, the AOS of BMO-8 h, BMO-24 h, BMO-48 h 2 3 2 (pyrolusite) were used to calibrate AOS as a function of edge was 3.6 ± 0.3 and the AOS of Todorokite-STD was energy. 3.7 ± 0.3. A double-hump in the range from 6540 to 6545 eV, arising from bound state quadrupole-allowed 1s to 3d transitions (Saratovsky et al., 2006), was found in tively compare BMO, BMO-Mg and the refluxed products the pre-edge region of the X-ray absorption spectrum of with todorokite. The k space and R space EXAFS spectrum the biogenic Mn oxide, each refluxed product and todorok- for BMO, as shown in Fig. 3, are consistent with those pub- ite (Figs. 1 and 2), suggesting that these Mn cations are lished for the biogenic Mn oxides produced by other model octahedrally coordinated. Therefore, the refluxed products Mn oxidizing bacteria, such as P. putida strain MnB1 (Vill- of BMO-Mg consist primarily of Mn(IV), with a fraction alobos et al., 2003, 2006), L. discophora SP6 (Kim et al., of Mn at lower valences, and exhibit octahedral coordina- 2003; Saratovsky et al., 2006) and Bacillus sp. strain SG-1 tion geometry, common to phyllomanganates and (Webb et al., 2005a,b, 2006). tectomanganates. The region of the EXAFS spectrum from 7.5 to 9.5 A˚ 1 varies the most between tunnel and layer manganate struc- 3.1.2. EXAFS tures (McKeown and Post, 2001; Manceau et al., 2005; EXAFS spectroscopy probes the average local coordina- Webb et al., 2006). After the reflux treatment, this region tion environment around Mn to approximately 6 A˚ (Vill- shows the most dramatic changes (Fig. 3a). Instead of dis- alobos et al., 2003, 2006; Webb et al., 2005a,b, 2006; tinct peaks, two steadily rising slopes gradually appear be- Saratovsky et al., 2006, 2009) and was used to quantita- tween 7.5 and 9.5 A˚ 1, indicative of the formation of a 3236 X.H. Feng et al. / Geochimica et Cosmochimica Acta 74 (2010) 3232–3245

(McKeown and Post, 2001; Kim et al., 2003). Furthermore, the ﬁtting parameters, including focc, b, and the number of corner sharing Mn, which are sensitive to the size of the tunnel (Webb et al., 2005a), for the reﬂuxed products, BMO-24 h and BMO-48 h, are similar to those of Todorok- ite-STD (Table 1).

3.2. SR-XRD

SR-XRD analysis was used to investigate the evolution of the long-range structure of the biogenic MnOx during the refluxing process. Diffraction patterns of BMO and BMO-Mg are similar, and exhibit two broad peaks at 0.246 and 0.142 nm resulting from the reflection of the ab-layer plane (Villalobos et al., 2003, 2006; Saratovsky et al., 2006) of the biogenic Mn oxides (Fig. 5). The absence of a basal reflection peak for BMO and BMO-Mg (0.7 or 1.0 nm) demonstrates the poorly ordered stacking of adjacent layers, or discrete layers as in the case of d-MnO2 (Villalobos et al., 2006; Bodei et al., 2007). Diffrac- tion patterns of the refluxed products have broad peaks centered at 0.477, 0.248, 0.154, 0.146 and 0.142 nm, all of which can be attributed to the monoclinic todorokite structure (JCPDS 38-475). The peak at 0.24 nm has a similar composite shape to that for todorokite and is diagnostic for todorokite, differentiating todorokite from other phyllomanganates with a similar basal plane reflections (1.0 nm), such as 1 nm vernadite (Bodei et al., 2007).

3.3. Electron microscopy

3 Fig. 3. Mn K-edge k -weighted EXAFS (a) and Fourier trans- FEG-SEM images (Fig. 6a) of the biogenic oxide show formed EXAFS (b) spectra (solid line) with best fit overlaid (dotted stringy, rope-like features, indicative of membrane bound line) from the full multiple scattering manganese EXAFS model (Webb et al., 2005a) for Pseudomonas putida GB-1 bacterial cell EPS desiccation during FEG-SEM sample preparation oxidation system and the intermediate products after Mg2+ (Toner et al., 2005). The morphology of the poorly crystalline exchange and reflux treatment for the different times as well as biogenic oxide particles consists of fibrillar thin planes with the synthesized todorokite standard. Note the change in diagnostic dimensions of 10 nm wide by 100 nm long surrounding features in the k space region between k 7–11 A˚ 1 which is the oblate spheroid shape of bacterial cells (Fig. 6b and c), highlighted by the vertical lines. similar to that reported for Mn oxides produced by L. discophora SP6 (Saratovsky et al., 2006) and P. putida strain MnB1 (Villalobos et al., 2003). After refluxing, the Mn tunnel structure manganate during refluxing (Webb et al., oxide particles maintained similar fibrillar morphology with 2005a, 2006; Bodei et al., 2007). In contrast to BMO-Mg, slightly smaller dimensions resulting from partial reductive the Fourier transform of the EXAFS spectra of the refluxed dissolution by residual biological substances. This indicates products gradually reveals typical characteristics of trans- a topotactic transformation process, in which the product formation of phyllomanganate to tectomanganate (Webb preserves the features of the precursor in morphology and et al., 2005a). These characteristics include a decrease in crystallization, as is the case for todorokite formation from the relative amplitude of peaks for the first shell edge-shar- phyllomanganate precursors (Bodei et al., 2007). The Mn ing Mn and the Mn–Mn multiple scattering over the reflux- oxide particles also became more closely associated with ing time from 8 to 48 h (Fig. 3b). Quantitative fitting the bacterial cells, which became slightly distorted during parameters (Table 1), using the full multiple scattering refluxing (Fig. 6d and e). In comparison, the surface of Mn EXAFS model (Webb et al., 2005a), confirmed a de- BMO-48 h became smooth, lacking the original biofilm, crease in Mn site occupancy (focc), an increase in dihedral which may have become mixed with the Mn oxide particles angle from out-of-plane bending (b), and an increase in the during refluxing. An HR-TEM image of an individual Mn number of corner sharing Mn octahedra due to formation oxide fiber of BMO-48 h shows the lattice fringes with a of a tunnel structure Mn oxide (Webb et al., 2005a, common dimension of 1 nm along the a* direction 2006). Mn EXAFS spectra of the refluxed products are (Fig. 6f), indicative of a tunnel width of three MnO6 octa- most similar to T(3 3) tunnel structure manganate (i.e., hedra as found in todorokite (Turner and Buseck, 1981; todorokite) rather than cryptomelane (T(2 2)), pyrolusite Post and Bish, 1988; Post, 1999). Furthermore, 0.5 nm (T(1 1)) (Fig. 4a and b) or psilomelane (T(2 3)) lattice fringes in the a* direction were also observed, Formation of todorokite from biogenic Mn oxides 3237

Table 1 Summary of EXAFS ﬁtting parameters from the Mn K-edge using the full multiple scattering Mn oxide model (Webb et al., 2005a) for the biogenic oxide, the products after diﬀerent treatments and the todorokite standard. Sample Ra v2b focc b (a-axis) b (b-axis) Shellc CNd Diat (A˚ ) r2 BMO 0.0226 2393.1 0.73(3) 0(3) 5(5) Mn–O 4 1.85(2) 0.006(4) Mn–O 2 1.93(1) 0.002(4) Mn–Mn edge 2 2.81(1) 0.005(2) Mn–Mn edge 4 2.87(3) Mn–O 4 3.47(6) 0.001(7) Mn–O 2 3.65(1) Mn–Mn corner 0.9(3) 3.51(2) 0.010(3) Mn–Na interlyr 1.0(5) 4.11(3) 0.004(3) Mn–O 4 4.67(2) 0.002(1) Mn–O 8 4.81(8) Mn–Mn diag 4 5.00(1) 0.005(1) Mn–Mn diag 2 5.20(2) Mn–Mn next 2 5.53(3) 0.006(2) Mn–Mn next 4 5.85(2) BMO-Mg 0.019 2889.8 0.76(3) 0(3) 8(3) Mn–O 4 1.85(1) 0.004(1) Mn–O 2 1.95(1) 0.001(1) Mn–Mn edge 2 2.81(1) 0.006(1) Mn–Mn edge 4 2.88(4) Mn–O 4 3.46(1) 0.001(1) Mn–O 2 3.69(2) Mn–Mn corner 0.5(5) 3.62(4) 0.009(1) Mn–Mg interlyr 0.5(5) 4.11(3) 0.001(6) Mn–O 4 4.68(3) 0.002(2) Mn–O 8 4.82(2) Mn–Mn diag 4 5.01(1) 0.004(2) Mn–Mn diag 2 5.20(4) Mn–Mn next 2 5.54(3) 0.006(2) Mn–Mn next 4 5.83(2) BMO-8 h 0.028 5077.6 0.68(3) 5(2) 16(4) Mn–O 4 1.86(1) 0.004(1) Mn–O 2 1.95(1) 0.001(1) Mn–Mn edge 2 2.85(2) 0.007(1) Mn–Mn edge 4 2.88(1) Mn–O 4 3.49(0.12) 0.001(1) Mn–O 2 3.64(4) Mn–Mn corner 1.4(8) 3.42(2) 0.009(4) Mn–Mg interlyr 0.3(3) 4.01(5) 0.001(3) Mn–O 4 4.62(3) 0.011(2) Mn–O 8 4.73(2) Mn–Mn diag 4 4.92(3) 0.006(2) Mn–Mn diag 2 5.06(6) Mn–Mn next 2 5.53(8) 0.009(4) Mn–Mn next 4 5.62(4) BMO-24 h 0.033 16916.1 0.57(3) 6(3) 22(6) Mn–O 4 1.86(1) 0.004(1) Mn–O 2 1.95(1) 0.001(1) Mn–Mn edge 2 2.86(2) 0.006(1) Mn–Mn edge 4 2.87(1) Mn–O 4 3.46(1) 0.001(1) Mn–O 2 3.59(2) Mn–Mn corner 2.1(1.1) 3.37(2) 0.012(5) Mn–Mg interlyr 0.5(0.3) 4.00(3) 0.001(2) Mn–O 4 4.61(8) 0.011(6) Mn–O 8 4.74(4) Mn–Mn diag 4 4.93(4) 0.007(5) Mn–Mn diag 2 5.07(0.10) Mn–Mn next 2 5.73(0.13) 0.007(6) Mn–Mn next 4 5.59(5) BMO-48 h 0.038 17923.2 0.53(3) 5(4) 22(4) Mn–O 4 1.86(1) 0.004(1) Mn–O 2 1.96(1) 0.001(1) Mn–Mn edge 2 2.87(2) 0.006(1) (continued on next page) 3238 X.H. Feng et al. / Geochimica et Cosmochimica Acta 74 (2010) 3232–3245

Table 1 (continued) Sample Ra v2b focc b (a-axis) b (b-axis) Shellc CNd Diat (A˚ ) r2 Mn–Mn edge 4 2.87(1) Mn–O 4 3.46(1) 0.001(1) Mn–O 2 3.59(3) Mn–Mn corner 2.5(1.3) 3.36(2) 0.012(5) Mn–Mg interlyr 0.6(0.3) 4.01(3) 0.001(3) Mn–O 4 4.61(8) 0.011(6) Mn–O 8 4.74(4) Mn–Mn diag 4 4.95(4) 0.007(5) Mn–Mn diag 2 5.08(0.10) Mn–Mn next 2 5.73(0.13) 0.007(7) Mn–Mn next 4 5.59(5) Todorokite-STD 0.017 6614.8 0.55(2) 6(3) 19(3) Mn–O 4 1.85(1) 0.004(1) Mn–O 2 1.96(1) 0.001(1) Mn–Mn edge 2 2.88(2) 0.007(1) Mn–Mn edge 4 2.86(1) Mn–O 4 3.45(1) 0.001(1) Mn–O 2 3.56(2) Mn–Mn corner 1.9(0.4) 3.38(1) 0.009(2) Mn–Mg interlyr 0.5(0.2) 3.96(2) 0.001(2) Mn–O 4 4.60(6) 0.009(4) Mn–O 8 4.77(3) Mn–Mn diag 4 4.98(2) 0.009(2) Mn–Mn diag 2 5.00(5) Mn–Mn next 2 5.70(5) 0.007(3) Mn–Mn next 4 5.57(2)

a R factor. b Chi squared. c Mn–Mn edge, Mn–Mn corner, Mn–Mn diag, Mn–Mn next corner sharing Mn shell denote edge-sharing Mn shell, corner sharing Mn shell, diagonal Mn shell and next to diagonal Mn shell, respectively. Mn–Mg(Na) interlayer denotes interlayer Mg(Na) shell (Webb et al., 2005a). d Fixed except for CN of Mn–Mn corner, Mn–Na interlayer and Mn–Mg interlayer. probably due to the intergrowth of a narrow tunnel with and Gru¨tter, 1956). Although todorokite formation by one MnO6 octahedron width, which is very common in the microbial mediation was reported by Takematsu et al. structure of natural and synthetic todorokite (Chukhrov (1984, 1988), the identification of the product via XRD et al., 1979; Turner et al., 1982; Golden et al., 1986; Feng should be viewed with skepticism. The product is likely et al., 2004; Bodei et al., 2007), or due to the orderly stack- 10-A˚ phyllomanganate or buserite because it was trans- ing of tunnel cations, Mg2+ or Mn2+, in the center of a tun- formed into birnessite with time. XAFS (XANES and EX- nel with a width of three MnO6 octahedra. AFS) spectroscopy, sensitive to local structure features, is a promising technique for the identification of biogenic origin 4. DISCUSSION Mn oxides. However, even with the aid of XAFS spectroscopy the identification can be ambiguous and erroneous 4.1. Structural transformation of the biogenic Mn oxide after such as in Kim et al. (2003). With careful comparison both reflux treatment in k and R space spectroscopy and fitting of EXAFS spectra the identification can be conclusive (Bargar et al., 2005, Due to the low degree of crystallinity, nanometer dimen- 2009; Webb et al., 2005a,b; Saratovsky et al., 2006, 2009; sions and similar edge-sharing MnO6 octahedral units that Villalobos et al., 2006). are arrayed in layers, as most layer and tunnel Mn oxides, In this paper, the comparison of spectra both in k and in structural characterization and identification of biogenic R space among the original biogenic Mn oxide, the refluxed Mn oxides is often ambiguous when based solely on products and tunnel structure Mn oxides with different tun- XRD analyses. For example, XRD patterns of biogenic nel sizes was conducted. The original biogenic Mn oxide Mn oxide, vernadite, buserite and todorokite are often (BMO) possesses common structural features of primary indistinguishable and are prone to be confused in their biological oxidation products of Mn oxidizing bacteria identification due to their similar diffraction features (Burns and is similar to d-MnO2 or hexagonal birnessite based et al., 1983, 1985; Giovanoli, 1985; Bodei et al., 2007; Sar- on its EXAFS spectral features, i.e., sharp peaks approxi- atovsky et al., 2009). Thus it can be explained why 10-A˚ mately at 8 and 9.2 A˚ 1 in k space in the EXAFS spectrum vernadite, buserite and todorokite were initially regarded (Fig. 3a), the high amplitude of peaks for the first shell as the same Mn oxide phase named as 10-A˚ manganite edge-sharing Mn (2.87 A˚ ) and the Mn–Mn multiple scatter- according to the d space of basal plane diffraction (Buser ing at about 5.6 A˚ in the Fourier transform of the EXAFS Formation of todorokite from biogenic Mn oxides 3239

Fig. 5. Synchrotron X-ray diffraction patterns for the Pseudomo- nas putida GB-1 bacterial cell oxidation system and the intermediate products after Mg2+ exchange and reflux treatment for the different times compared to the XRD pattern of the synthesized todorokite standard.

patterns (Fig. 5), which is attributed to thin or disordered layer stacking along the c* axis and the lack of c* periodicity (Villalobos et al., 2006; Bodei et al., 2007; Bargar et al., 2009), as in case of the precursor BMO and BMO-Mg.

4.2. Comparison of the transformation of chemically synthesized birnessite to todorokite and natural todorokite 3 Fig. 4. Comparison of Mn K-edge k -weighted EXAFS (a) and formation Fourier transformed EXAFS (b) between BMO-24 h and different tunnel structured Mn oxide minerals [todorokite (T(3 3)), Natural todorokites typically occur as poorly crystal- cryptomelane (T(2 2)) and pyrolusite (T(1 1))]. lized nano-particles mixed with many other minerals. Syn- pffiffiffi thesis of todorokite is a promising alternative to obtain spectra (Fig. 3b). The d(200, 110)/d(020, 310) ratio is 3 todorokite samples for fundamental research and industrial (Fig. 5), which also exemplifies the hexagonal symmetry application studies. Todorokite was first synthesized of BMO (Drits et al., 1997; Villalobos et al., 2003, 2006). through autoclave treatment of 10-A˚ Mg2+ exchanged bir- The close similarity between BMO and BMO-Mg EXAFS nessite (Mg-buserite) at 155 °C(Golden et al., 1986, 1987). spectra suggests that Mg2+ exchange does not change the Another thermally stable version of todorokite can be syn- local structure. thesized by a similar hydrothermal treatment via oxidation It is typically easy to tell todorokite from other tunnel of Mn(OH)2 with Mg(MnO4)2 to prepare precursor birnes- Mn oxides, but more difficult to distinguish between disor- site in alkali media (Shen et al., 1993). Other hydrothermal dered todorokite and disordered phyllomanganates. The methods for todorokite syntheses that use different birnes- comparison was also performed between the refluxed prod- site preparation procedures or microwave heating have also ucts and different phyllomanganates, such as d-MnO2, hex- been reported (Feng et al., 1995, 1998; Vileno et al., 1998; agonal birnessite, random stacked birnessite (RSB) and Ching et al., 1999; Luo et al., 1999; Liu et al., 2005). In triclinic birnessite (Fig. 7). From the comparison both in addition to hydrothermal methods, a mild reflux procedure k and in R space the product is different from any of the under atmospheric pressure was proposed to synthesize phyllomanganites. It can be seen that the spectra of the re- todorokite in our previous work (Feng et al., 2004). There- fluxed products and the EXAFS fitting results (Fig. 3 and fore, preparation of Na-birnessite, exchange of 7-A˚ Na-bir- Table 1), especially the results of vacancy site number nessite with a cation (generally Mg2+ is used) to obtain a (focc), corner sharing Mn number and out-of-plane bend- 10-A˚ buserite, and then hydrothermal or reflux treatment ing angle (b) match very well with those of todorokite. of 10-A˚ buserite at a relatively high temperature are com- Therefore, after the reflux treatment, Mg exchanged bio- mon steps in todorokite syntheses. genic Mn oxide gradually transformed into a tunnel struc- The reflux treatments used to transform biogenic Mn ture manganate similar to todorokite. oxide into todorokite are generally similar to those used SR-XRD analyses further confirmed the transformation to transform chemically synthesized birnessite into tod- at long-range structural scale. There is an absence of the ba- orokite. However, the precursors are different, not only is sal plane reflection (1.0 nm) in the refluxed products XRD our precursor biological in origin, without a basal plane dif- 3240 X.H. Feng et al. / Geochimica et Cosmochimica Acta 74 (2010) 3232–3245 fraction peak, but it also possesses hexagonal symmetry, ment, instead it converts to cryptomelane, ramsdellite and distinct to other precursors which possess triclinic symme- another unidentifiable phase (Fig. EA-1). Attempts to syn- try. Interestingly, in all reported todorokite synthesis proce- thesize todorokite from 10-A˚ vernadite, with hexagonal dures, the precursor minerals are exclusively triclinic symmetry, were also unsuccessful (Bodei et al., 2007). birnessite, which forms in alkali media and has one third Topotactic transformation from layer Mn oxides to of Mn(III)O6 octahedra in MnO6 octahedral unit (Drits todorokite was observed in both synthesized and natural et al., 1997; Lanson et al., 2002). todorokites (Chukhrov et al., 1979; Golden et al., 1986, Microbially mediated Mn(II) oxidation is believed to be 1987; Shen et al., 1993; Feng et al., 2004; Bodei et al., the dominant source of Mn oxides in marine, freshwater 2007). In these cases, todorokites formed in situ from pre- and subsurface aquatic environments (Tebo et al., 2004; cursor phyllomanganates, exhibited a morphology of platy Webb et al., 2006; Bargar et al., 2009). The primary prod- matrix consisting of twinned fibers matted at 120° from ucts of microbial Mn(II) oxidation are poorly crystallized one another, from which fiber crystals extended. The layer Mn oxides with a hexagonal symmetry, similar to HR-TEM results indicate the topotactic transformation d-MnO2 (Tebo et al., 2004; Bargar et al., 2005, 2009; Webb from biogenic Mn oxide to todorokite-like phase after et al., 2005a,b; Miyata et al., 2006; Saratovsky et al., 2006; refluxing. While we do not observe the twinned fibers in Villalobos et al., 2006). Therefore, natural todorokites in the refluxed products, this possibly results from the the environment are considered to form from layer struc- weakly crystallized BMO precursor with a low degree of tured Mn oxides with hexagonal symmetry, such as verna- 3D periodicity (Bodei et al., 2007). Due to the topotactic dite (d-MnO2), which are generally of biogenic origin (Tebo transformation from biogenic Mn oxide, the refluxed et al., 2004; Webb et al., 2006; Bodei et al., 2007; Bargar products have thin layers along the c* axis and lack c* et al., 2009). However, this transformation process has periodicity. In addition because the 1 nm basal plane not been documented in the laboratory. Acid birnessite, reflection is a doublet of the (1 0 0) and (0 0 1), the lack synthesized in acidic media with a hexagonal symmetry, of 1 nm reflection indicated that the refluxed products also does not transform into todorokite through the same treat- lack a* periodicity. Intergrowth of T(3 n) building

a b d e

Mn oxide

Mn oxide Biofilm Cell Cell

c f

Fig. 6. Electron micrographs of the biogenic Mn oxide produced by Pseudomonas putida GB-1 (a–c) and the product (d–f) after Mg2+ exchange and reflux treatment for 48 h: (a) FEG-SEM image of the biogenic Mn oxide showing the cells, Mn oxide particles and associated desiccated EPS, (b) TEM image of the biogenic Mn oxide shows the filament or fiber like biogenic Mn oxide surrounding the bacterial cells, (c) HR-TEM of the biogenic Mn oxide shows long and thin sheet morphologies (50 by 5 nm), (d) FEG-SEM image of the refluxing product shows more closely the association of cells and the Mn oxide particles, (e) TEM image of the refluxing product shows the slightly distorted cells and the peripheral fibrous Mn oxide, (f) HR-TEM image of an individual fiber of the refluxing product shows 1 and 0.5 nm spacings in the a direction. Formation of todorokite from biogenic Mn oxides 3241 blocks with different tunnel sizes along a* in the natural The poorly crystalline todorokite-like phase formed and synthetic todorokite samples is quite common from biogenic Mn oxides may progressively increase in size (Chukhrov et al., 1979; Turner et al., 1982; Golden and crystallinity to become the final todorokite products et al., 1986; Feng et al., 2004; Bodei et al., 2007). The under conditions similar to natural todorokite formation poorly ordered layer Mn oxide precursor would cause described in Buatier et al. (2004). Crystallinity was also ob- more incoherent tunnel width of the formed todorokites served to increase in the process of natural todorokite for- (Bodei et al., 2007). Accordingly, the structural character- mation in hemipelagic sediments (Bodei et al., 2007). In istics of biogenic Mn oxide precursor may cause disorder addition, todorokite formed under reflux conditions exhib- array along a* in the refluxed products. From the HR- ited increasing XRD peak intensities and crystallinity with TEM images (Fig. 6), different spacings in the a* direction reflux time (Feng et al., 2004). However, in this case, the were observed. Such morphologic features of variable tun- 0.477 and 0.248 nm XRD peaks diagnostic for todorokite nel sizes along a* were typical and could be observed in decrease in intensity with the reflux time (Fig. 5). This the whole fiber of the refluxed products. The lattice fringes can be ascribed to partial reductive dissolution and the are not very clear due to the low crystallinity and thin resultant thinner layer of the products as the reflux time in- layer along c*. We were unsuccessful in acquiring better creased. The analyses of the reflux solution indicates that images with high resolution and magnification under the Mn(II) concentration increased from 0 to 20 mg/L HR-TEM because the products were slightly sensitive to and pH increased from 5.1 to above 7.0 when reflux pro- the electron beam. The lattice fringes with different spac- ceeded from 0 to 48 h (Table EA-1). Therefore, the pure ing in the a* direction indicate that tunnel structures were Mn oxide system and/or high content of residual biological formed in the products. Thus, the transformation of the substances relative to the natural environment and previous biogenic oxide to a todorokite-like phase after refluxing reflux system may be the reason that XRD peaks slightly can be illustrated in Fig. 8. attenuate with the increasing reflux time.

4.3. Implications for the formation of bio-related origin of natural todorokites

Based on data from laboratory and ﬁeld observations, todorokite does not form directly via a precipitation process (Siegel and Turner, 1983; Golden et al., 1986; Shen et al., 1993; Feng et al., 1995, 1998, 2004; Buatier et al., 2004; Bodei et al., 2007), instead it is likely a product from the transformation of a 10-A˚ phyllomanganate precursor

3 3 a* b*

c* 5 4 3 3 3 a*

Fig. 8. Diagram of transformation of the biogenic Mn oxide to a todorokite-like phase. The biogenic Mn oxide and the Mg2+ exchanged product were thin layer structured phyllomanganates 3 Fig. 7. Mn K-edge k -weighted EXAFS (a) and Fourier transfor- which lack c* periodicity with interlayer spacing of three MnO6 mation EXAFS (b) comparison of BMO-24 h and BMO with octahedra width along c*. After refluxing, thin tunnel structured different phyllomanganites (d-MnO2, hexagonal birnessite, random todorokite-like phases, which lack c* periodicity with interlayer stacked birnessite and triclinic birnessite). Note the change in spacing of three MnO6 octahedra width along c*, and intergrowth diagnostic features in the k space region between k 7–11 A˚ 1 with different spacings along a*, were formed through topotactic which is highlighted by the vertical lines. transformation. 3242 X.H. Feng et al. / Geochimica et Cosmochimica Acta 74 (2010) 3232–3245

(Bodei et al., 2007). Recently, a todorokite-like tunnel failures to obtain todorokite from acid birnessite and 10- ˚ structural MnOx formed by the fungus Acremonium sp. A vernadite, as discussed above, can be explained by lack Strain KR21-2 was reported (Saratovsky et al., 2009). If of enough Mn(III) and hexagonal symmetry in their the above assumption is correct, this fungally mediated MnO6 layer. Furthermore, the kinetics of todorokite trans- MnOx material may not yield from the primary oxidation formation from biogenic Mn oxides can be increased by of Mn2+, but is a secondary product. The association of refluxing, which is another reason why todorokite-like biogenic Mn oxide with todorokite, or bio-related origin phases formed in this study. At the temperature of reflux of todorokite, was observed in many natural environments (100 °C), Mg2+-buserite completely converts to todorokite (Burns and Burns, 1978b; Siegel and Turner, 1983; Buatier within 8 h (Feng et al., 2004; Cui et al., 2006). It takes et al., 2004; Bodei et al., 2007). Therefore, the simple bio-re- 48 h and 120 h for the similar conversion at 90 °C and lated formation pathway of natural todorokite can be writ- 80 °C, respectively. When the temperature is lowered to ten as: biogenic Mn oxide ? 10-A˚ phyllomanganate ? 40 °C, only part of Mg2+-buserite converts to todorokite todorokite. The results of this study provide for the first even after aging for 35 days. At relatively low temperatures, time the experimental evidence of biogenic Mn oxide trans- the rate of todorokite formation decreases sharply (Cui formation to a todorokite-like phase. et al., 2006), explaining why todorokite tends to occur prev- Stable 10-A˚ phyllomanganate is the prerequisite for tod- alently in ocean hydrothermal deposits over diagenetic orokite formation because its interlayer space matches the deposits. Conversely, 10-A˚ phyllomanganate prevails in T(3 3) tunnel size (Burns et al., 1983; Post and Bish, diagenetic deposits (Bodei et al., 2007). Consequently, sta- 1988). Thus interlayer cations with a high enthalpy of ble 10-A˚ phyllomanganate, appropriate Mn(III) content hydration, such as Mg2+,Cu2+,Ni2+ and Ca2+, are needed and/or triclinic symmetry of the phyllomanganate and rel- to form stable 10-A˚ phyllomanganate before the todorokite atively high temperature conditions are the three key fac- formation. BMO without exchanged Mg2+ cannot be con- tors determining todorokite formation. Marine or verted to as todorokite phase via reflux, but a T(2 2) terrestrial environments which meet with such conditions, cryptomelane-like phase forms instead (Fig. EA-2). In addi- such as marine hydrothermal Mn deposits (Usui et al., tion, as proposed by Bodei et al. (2007), the ideal precursor 1989, 1997; Buatier et al., 2004; Bodei et al., 2007; Takah- for defect-free T(3 3) todorokite is a phyllomanganate ashi et al., 2007; Dubinin et al., 2008), marine diagenetic with high Mn(III); low Mn(III) phyllomanganate results Mn concretions (Yoshikawa, 1991; Usui et al., 1997; in a highly defective todorokite. The redox conditions, Takahashi et al., 2007), and soils or sediments enriched influencing the Mn(III) content in the phyllomanganate, with Ca (Taylor et al., 1964; Turner and Buseck, 1981; also have effects on the tunnel size of natural todorokite McKenzie, 1989; Bilinski et al., 2002; Tan et al., 2006; Man- along a*(Mellin and Lei, 1993; Lei, 1996). ceau et al., 2007), are expected to favor todorokite Our recent work also indicates that Mn(III) plays a key formation. role in the transformation of triclinic layered Na-buserite to A two-step dissolution-recrystallization process of natu- todorokite via reflux treatment; the transformation from ral todorokite formation in ocean nodules was proposed by Na-buserite to todorokite decreased gradually with decreas- Burns and Burns (1978b) as follows. First, Mn oxides ing Mn(III) content (Cui et al., 2008, 2009b). The elonga- formed at the surface of detrital matter were partially dis- tion and weakening of the Mn(III)–O bond in the solved by the surrounding organic substances releasing octahedral layer due to the Jahn-Teller effect will cause Mn2+; second, the cryptocrystalline Mn oxide of biogenic the kink-like fold of the layer (Bodei et al., 2007), from or abiotic origin, adsorbs the released Mn2+ and transforms which the tunnel walls are constructed. Therefore, in this into todorokite. Bodei et al. (2007) suggested a three-step system, some Mn(III) may be produced by re-oxidation process of the phyllomanganate to todorokite conversion of Mn(II) originating from the partial reductive dissolution in ocean sediments, which involves another step of semi- of biogenic Mn oxide, which favors todorokite formation. ordered 10-A˚ phyllomanganate formation before the Such Mn(III) production could also account for increased conversion. The process of todorokite-like phase formation Mn(III) in the refluxed product than its precursor BMO in this study verifies the above assumed processes of tod- or Todorokite-STD. The symmetry can be changed from orokite formation in the marine environment. It should be triclinic to hexagonal as the Mn(III) content of triclinic bir- pointed out that the biogenic Mn oxide precursor used in nessite decreases in the layer (Drits et al., 1997; Silvester this study was produced by a freshwater bacterium, i.e., P. et al., 1997). On the contrary, it is reasonable to infer that putida strain GB-1, but it will not influence the application hexagonal phyllomanganates could be converted to triclinic of the implications or this work to marine environments ones with an increase of Mn(III), as the case in this work. due to the similar primary products of microbial Mn(II) The decrease of amplitude in the Mn–Mn multiple scatter- oxidation in terms of structure, morphology and crystallin- ing peak of 5–6 A˚ in the R space EXAFS spectra (Fig. 3b) ity either by marine bacteria or freshwater bacteria. and the increase in dihedral angle from out-of-plane bending (Table 1) after reflux also make the inference plausible. 5. SUMMARY AND CONCLUSION Thus, the above todorokite formation pathway can be spec- ified as: hexagonal biogenic Mn oxide ? 10-A˚ triclinic The formation pathway of todorokite from layer struc- phyllomanganate ? todorokite. The more the structure of tured Mn oxides with hexagonal symmetry, biogenic Mn ˚ the precursors departs from 10-A triclinic phyllomanga- oxides or vernadite (d-MnO2), has been speculative due nate, the more defects the todorokite would exhibit. The to the lack of direct evidence (Burns and Burns, 1978b; Formation of todorokite from biogenic Mn oxides 3243

Mandernack et al., 1995; Post, 1999; Buatier et al., 2004; from Pinal Creek, Arizona, USA, and in hot-spring deposits Bodei et al., 2007). In the present study, it was discovered from Yuno-Taki falls, Hokkaido, Japan. Am. Mineral. 87, 580– that a nano-crystalline todorokite-like phase forms from 591. biogenic Mn oxides at atmospheric pressure via a refluxing Bodei S., Manceau A., Geoffroy N., Baronnet A. and Buatier M. process. This implies that natural todorokite in marine and (2007) Formation of todorokite from vernadite in Ni-rich hemipelagic sediments. Geochim. Cosmochim. Acta 71, 5698– terrestrial surface environments may originate from bio- 5716. genic Mn oxides and is subject to recrystallization from Boogerd F. C. and de Vrind J. P. M. (1987) Manganese oxidation its poorly crystallized form. The symmetry of the phyllo- by Leptothrix discophora. J. Bacteriol. 169, 489–494. manganates may convert from hexagonal to triclinic, as Buatier M. D., Guillaume D., Wheat C. G., Herve L. and Adatte determined by content of Mn(III) in the MnO6 layer, before T. (2004) Mineralogical characterization and genesis of hydro- their transformation to todorokite. This fundamental thermal Mn oxides from the flank of the Juan the Fuca Ridge. knowledge of biogenic Mn oxide transformation to tod- Am. Mineral. 89, 1807–1815. orokite is critical to understanding the origin of natural Burle E. and Kirby-Smith W. W. (1979) Application of formal- todorokite and the geochemistry of Mn (hydro-) oxide min- doxime colorimetric method for the determination of manga- erals in nature. Furthermore, if different exchangeable cat- nese in the pore water of anoxic estuarine sediments. Estuar. Coast. 2, 198–201. ions and solution conditions are present in the above Burns V. M. and Burns R. G. (1978a) Authigenic todorokite and experimental system, other Mn oxides may be expected to phillipsite inside deep-sea manganese nodules. Am. Mineral. 63, form from biogenic Mn oxide through the refluxing process 827–831. (Fig. EA-2). This process accelerates the rate of transforma- Burns V. M. and Burns R. G. (1978b) Post-depositional metal tion reactions of biogenic Mn oxides without apparent enrichment processes inside manganese nodules from the north damage or dissolution. This will facilitate our ability to equatorial Pacific. Earth Planet. Sci. Lett. 39, 341–348. investigate the relationship and underlying mechanisms Burns R. G., Burns V. M. and Stockman H. (1983) A review of the for biogenic Mn oxide transformation, a process commonly todorokite–buserite problem: implications to the mineralogy of occurring in ocean Mn deposits and soil ferromanganese marine manganese nodules. Am. Mineral. 68, 972–980. aggregates, and the biosynthesis of various porous OMS Burns R. G., Burns V. M. and Stockman H. (1985) The todorokite–buserite problem: further consideration. Am. Min- nano-crystallites. It is hoped that our results will help to eral. 68, 972–980. stimulate such investigations. Buser W. and Gru¨tter A. (1956) Ueber die Natur der Mangank- nollen. Schweiz. Mineral. Petrogr. Mitt. 36, 49–62. ACKNOWLEDGEMENTS Ching S., Krukowska K. S. and Suib S. L. (1999) A new synthetic route to todorokite-type manganese oxides. Inorg. Chim. Acta We thank S.M. Webb and K. Pandya for their technical sup- 294, 123–132. port with the SR-XRD and XAFS analyses. We are grateful to Chukhrov F. V., Gorshkov A., Sivtsov A. V. and Berezovskaya V. V. K. Czymmek, S. Modla and D. Powell for their help with the (1979) New data on natural todorokites. Nature 278, 631–632. FEG-SEM and TEM analyses. We thank Dr. Jeffrey E. Post Cui H. J., Feng X. H., He J. Z., Tan W. F. and Liu F. (2006) Effects (Smithsonian Institution, Washington, DC) for his insightful re- of reaction conditions on the formation of todorokite at view of the manuscript prior to submission. We gratefully acknowl- atmospheric pressure. Clays Clay Miner. 54, 605–615. edge the two anonymous reviewers for their critical and very Cui H. J., Liu X. W., Tan W. F., Feng X. H., Liu F. and Ruan H. helpful comments on the manuscript. X.F. Feng thanks the Natu- D. (2008) Influence of Mn(III) availability on the phase ral Science Foundation of China (Nos. 40830527 and 40971142), transformation from layered buserite to tunnel-structured Program for New Century Excellent Talents in University and todorokite. Clays Clay Miner. 56, 397–403. the Foundation for the Author of National Excellent Doctoral Dis- Cui H. J., Feng X. H., Tan W. F., He J. Z., Hu R. G. and Liu F. sertation of PR China (No. 200767) for financial support. (2009a) Synthesis of todorokite-type manganese oxide from Cu- buserite by controlling the pH at atmospheric pressure. Micropor. Mesopor. Mater. 117, 41–47. APPENDIX A. SUPPLEMENTARY DATA Cui H. J., Qiu G. H., Feng X. H., Tan W. F. and Liu F. (2009b) Birnessites with different average manganese oxidation states were synthesized, characterized, and transformed to todorokite Supplementary data associated with this article can be at atmospheric pressure. Clays Clay Miner. 57(6), 715–724. found, in the online version, at doi:10.1016/j.gca.2010.03. Dixon J. B. and Skinner H. C. W. (1992) Manganese minerals in 005. surface environments. In Biomineralization Processes of Iron and Manganese. Modern and Ancient Environments (eds. H. C. REFERENCES W. Skinner and R. W. Fitzpatrick). CATENA Verlag, Crem- lingen-Destedt, Germany. Bargar J. R., Tebo B. M., Bergmann U., Webb S. M., Glatzel P., Drits V. A., Silvester E., Gorshkov A. I. and Manceau A. (1997) Chiu V. Q. and Villalobos M. (2005) Biotic and abiotic The structure of synthetic monoclinic Na-rich birnessite and products of Mn(II) oxidation by spores of the marine Bacillus hexagonal birnessite. Part 1. Results from X-ray diffraction and sp. strain SG-1. Am. Mineral. 90, 143–154. selected area electron diffraction. Am. Mineral. 82, 946–961. Bargar J. R., Fuller C. C., Marcus M. A., Brearley A. J., De la Dubinin A., Uspenskaya T., Gavrilenko G. and Rashidov V. (2008) Rosa M. P., Webb S. M. and Caldwell W. A. (2009) Structural Geochemistry and genesis of Fe–Mn mineralization in island arcs characterization of terrestrial microbial Mn oxides from Pinal in the west Pacific Ocean. Geochem. Int. 46, 1206–1227. Creek, AZ. Geochim. Cosmochim. Acta 73, 889–910. Feng Q., Kanoh H., Miyai Y. and Ooi K. (1995) Metal ion Bilinski H., Giovanoli R., Usui A. and Hanzel D. (2002) extraction/insertion reactions with todorokite-type manganese Characterization of Mn oxides in cemented streambed crusts oxide in the aqueous phase. Chem. Mater. 7, 1722–1727. 3244 X.H. Feng et al. / Geochimica et Cosmochimica Acta 74 (2010) 3232–3245

Feng Q., Yanagisawa K. and Yamasaki N. (1998) Hydrothermal McKenzie R. M. (1971) The synthesis of birnessite, cryptomelane, soft chemical process for synthesis of manganese oxides with and some other oxides and hydroxides of manganese. Miner. tunnel structures. J. Porous Mater. 5, 153–161. Mag. 38, 493–503. Feng Q., Kanoh H. and Ooi K. (1999) Manganese oxide porous McKeown D. A. and Post J. E. (2001) Characterization of crystals. J. Mater. Chem. 9, 319–333. manganese oxide mineralogy in rock varnish and dendrites Feng X. H., Tan W. F., Liu F., Wang J. B. and Ruan H. D. (2004) using X-ray absorption spectroscopy. Am. Mineral. 86, 701– Synthesis of todorokite at atmospheric pressure. Chem. Mater. 713. 16, 4330–4336. Mellin T. A. and Lei G. (1993) Stabilization of 10 A˚ -manganates Feng X. H., Zhai L. M., Tan W. F., Liu F. and He J. Z. (2007) by interlayer cations and hydrothermal treatment: implications Adsorption and redox reactions of heavy metals on synthesized for the mineralogy of marine manganese concretions. Mar. Mn oxide minerals. Environ. Pollut. 147, 366–373. Geol. 115, 67–83. Gadde R. R. and Laitinen H. A. (1974) Studies of heavy metal Miyata N., Tani Y., Maruo K., Tsuno H., Sakata M. and Iwahori adsorption by hydrous iron and manganese oxides. Anal. Chem. K. (2006) Manganese(IV) oxide production by Acremonium sp. 46(13), 2022–2026. strain KR21-2 and extracellular Mn(II) oxidase activity. Appl. Glasby G. P. (2006) Manganese: predominant role of nodules and Environ. Microbiol. 72, 6467–6473. crusts. In Marine Geochemistry (eds. H. D. Schulz and M. Nealson K. H., Tebo B. M. and Rosson R. A. (1988) Occurrence Zabel). Springer, pp. 371–427. and mechanisms of microbial oxidation of manganese. Adv. Giovanoli R. (1985) A review of the todorokite–buserite problem: Appl. Microbiol. 33, 279–318. implications to the mineralogy of marine manganese nodules: Parikh S. J. and Chorover J. (2005) FTIR spectroscopic study of discussion. Am. Mineral. 70, 202–204. biogenic Mn-oxide formation by Pseudomonas putida GB-1. Goldberg E. D. (1954) Marine geochemistry I. Chemical scavengers Geomicrobiol. J. 22, 207–218. of the sea. J. Geol. 62, 249–265. Post J. E. and Bish D. L. (1988) Rietveld refinement of the Golden D. C., Chen C. C. and Dixon J. B. (1986) Synthesis of todorokite structure. Am. Mineral. 73, 861–869. todorokite. Science 231, 717–719. Post J. E. (1999) Manganese oxide minerals: crystal structures and Golden D. C., Chen C. C. and Dixon J. B. (1987) Transformation economic and environmental significance. Proc. Natl. Acad. Sci. of birnessite to buserite, todorokite, and manganite under mild USA 96, 3447–3454. hydrothermal treatment. Clays Clay Miner. 35, 271–280. Saratovsky I., Wightman P. G., Pasten P. A., Gaillard J. F. and Hammersley A. P., Svensson S. O., Han M., Fitch A. N. and Poeppelmeier K. R. (2006) Manganese oxides: parallels between Hausermann D. (1996) Two-dimensional detector software: abiotic and biotic structures. J. Am. Chem. Soc. 128, 11188– from real detector to idealised image or two-theta scan. High 11198. Press. Res. 14, 235–248. Saratovsky I., Gurr S. J. and Hayward M. A. (2009) The structure Kim H. S., Pasten P. A., Gaillard J. F. and Stair P. C. (2003) of manganese oxide formed by the fungus Acremonium sp. Nanocrystalline todorokite-like manganese oxide produced by strain KR21-2. Geochim. Cosmochim. Acta 73, 3291–3300. bacterial catalysis. J. Am. Chem. Soc. 125, 14284. Silvester E., Manceau A. and Drits V. A. (1997) The structure of Lanson B., Drits V. A., Feng Q. and Manceau A. (2002) Crystal synthetic monoclinic Na-rich birnessite and hexagonal birnes- structure determination of synthetic Na-rich birnessite: evi- site. Part 2. Results from chemical studies and EXAFS dence for a triclinic one-layer cell. Am. Mineral. 87, 1662– spectroscopy. Am. Mineral. 82, 962–978. 1671. Shen Y. F., Zerger R. P., DeGuzman R. N., Suib S. L., McCurdy Lei G. B. (1996) Crystal structures and metal uptake capacity of 10 L., Potter D. I. and O’Young C. L. (1993) Manganese oxide angstrom-manganates: an overview. Mar. Geol. 133, 103–112. octahedral molecular sieves: preparation, characterization and Liu Z. H., Kang L., Ooi K., Yoji M. and Feng Q. (2005) Studies on application. Science 260, 511–515. the formation of todorokite-type manganese oxide with differ- Siegel M. D. and Turner S. (1983) Crystalline todorokite associated ent crystalline birnessites by Mg2+-templating reaction. J. with biogenic debris in manganese nodules. Science 219, 172– Colloid Interface Sci. 285, 239–246. 174. Luo J., Zhang Q., Huang A., Giraldo O. and Suib S. L. (1999) Suib S. L. (2008) Structure, porosity, and redox in porous Double-aging method for preparation of stabilized Na-buserite manganese oxide octahedral layer and molecular sieve materi- and transformations to todorokites incorporated with various als. J. Mater. Chem. 18, 1623–1631. metals. Inorg. Chem. 38, 6106–6113. Tan W. F., Liu F., Li Y. H., Hu H. Q. and Huang Q. Y. (2006) Manceau A., Tommaseo C., Rihs S., Geoffroy N., Chateigner D., Elemental composition and geochemical characteristics of iron– Schlegel M., Tisserand D., Marcus M. A., Tamura N. and Chen manganese nodules in main soils of China. Pedosphere 16(1), Z. S. (2005) Natural speciation of Mn, Ni, and Zn at the 72–81. micrometer scale in clayey paddy soil using X-ray fluorescence, Taylor R. M., McKenzie R. M. and Norrish K. (1964) The absorption, and diffraction. Geochim. Cosmochim. Acta 69, mineralogy and chemistry of manganese in some Australian 4007–4034. soils. Aust. J. Soil Res. 2, 235–248. Manceau A., Kersten M., Marcus M. A., Geoffroy N. and Granina Takahashi Y., Manceau A., Geoffroy N., Marcus M. A. and Usui L. (2007) Ba and Ni speciation in a nodule of binary Mn oxide A. (2007) Chemical and structural control of the partitioning of phase composition from Lake Baikal. Geochim. Cosmochim. Co, Ce, and Pb in marine ferromanganese oxides. Geochim. Acta 71, 1967–1981. Cosmochim. Acta 71, 984–1008. Mandernack K. W., Post J. and Tebo B. M. (1995) Manganese Takematsu N., Sato Y. and Okabe S. (1984) The formation of mineral formation by bacterial spores of the marine Bacillus, todorokite and birnessite in sea water pumped from under strain SG-1: evidence for the direct oxidation of Mn(II) to ground. Geochim. Cosmochim. Acta 48, 1099–1106. Mn(IV). Geochim. Cosmochim. Acta 59, 4393–4408. Takematsu N., Kusakabe H., Sato Y. and Okabe S. (1988) McKenzie R. M. (1989) Manganese oxides and hydroxides. In Todorokite formation in seawater by microbial mediation. J. Minerals in Soil Environments, second ed. (eds. J. B. Dixon Oceanogr. Soc. Jpn. 44, 235–243. and S. B. Weed). SSSA Book Series No. 1. SSSA Inc. pp. Tebo B. M., Bargar J. R., Clement B. G., Dick G. J., Murray K. J., 439–465. Parker D., Verity R. and Webb S. M. (2004) Biogenic Formation of todorokite from biogenic Mn oxides 3245

manganese oxides: properties and mechanisms of formation. monas putida strain MnB1. Geochim. Cosmochim. Acta 67, Annu. Rev. Earth Planet. Sci. 32, 287–328. 2649–2662. Toner B., Fakra S., Villalobos M., Warwick T. and Sposito G. Webb S. M. (2005) SIXPACK: a graphical user interface for (2005) Spatially resolved characterization of biogenic manga- XAS analysis using IFEFFIT. Phys. Scr. T115, 1011– nese oxide production within a bacterial biofilm. Appl. Environ. 1014. Microbiol. 71, 1300–1310. Webb S. M., Fuller C. C., Tebo B. M. and Bargar J. R. (2006) Turner S. and Buseck P. R. (1981) Todorokites: a new family of Determination of uranyl incorporation into biogenic manga- naturally occurring manganese oxides. Science 212, 1024–1027. nese oxides using X-ray absorption spectroscopy and scattering. Turner S., Siegel M. D. and Buseck P. R. (1982) Structural features Environ. Sci. Technol. 40, 771–777. of todorokite intergrowths in manganese nodules. Nature 296, Webb S. M., Tebo B. M. and Bargar J. R. (2005a) Structural 841–842. characterization of biogenic manganese oxides produced in Usui A., Mellin T. A., Nohara M. and Yuasa M. (1989) Structural seawater by the marine Bacillus sp. strain SG-1. Am. Mineral. stability of marine 10 A˚ manganates from the Ogasawara 90, 1342–1357. (Bonin) arc: implication for low-temperature hydrothermal Webb S. M., Tebo B. M. and Bargar J. R. (2005b) Structure activity. Mar. Geol. 86, 41–56. influence of sodium and calcium ions on the biogenic manga- Usui A., Bau M. and Yamazaki T. (1997) Manganese microchim- nese oxides produced by the marine Bacillus sp. strain SG-1. neys buried in the Central Pacific pelagic sediments: evidence of Geomicrobiol. J. 22, 181–193. intraplate water circulation? Mar. Geol. 141, 269–285. Yang D. S. and Wang M. K. (2002) Syntheses and characterization Vileno E., Ma Y., Zhou H. and Suib S. L. (1998) Facile synthesis of of birnessite by oxidizing pyrochroite in alkaline conditions. synthetic todorokite (OMS-1), co-precipitation reactions in the Clays Clay Miner. 50, 63–69. presence of a microwave field. Micropor. Mesopor. Mater. 20, Yoshikawa K. (1991) The relationship between manganese miner- 3–15. als and metallic elements in deep-sea manganese nodules. Mar. Villalobos M., Lanson B., Manceau A., Toner B. and Sposito G. Geol. 101, 267–286. (2006) Structural model for the biogenic Mn oxide produced by Pseudomonas putida. Am. Mineral. 91, 489–502. Associate editor: Jon Chorover Villalobos M., Toner B., Bargar J. and Sposito G. (2003) Characterization of the manganese oxide produced by Pseudo-